AUTOMATED RECORDING OF SOCIAL AND COGNITIVE BEHAVIORS IN GROUP-HOUSED COMMON MARMOSETS

Abstract Background The common marmosets (Callithrix jacchus) exhibit a wide repertoire of social behaviors and cognitive functions valuable for translational neuropsychiatric research. However, they are prone to stress, and environmental factors such as poor housing conditions and variations in the experimenter or handling methods lead to unintended reactions. Manual evaluation also restricts the types and quantities of measurable indices. Aims & Objectives We aimed to develop a method that can reliably capture a diverse array of behaviors in common marmosets. For this, our objectives were to: (1) Implement behavioral recording within a socially and physically enriched environment to facilitate the expression of a typical and rich behavioral repertoire, and (2) Automate data recording to minimize experimenter- animal contact and efficiently collect extensive data over an extended period, for multiple animals. Method Radio-frequency identification (RFID), a contactless identification technology, was used to continuously record the behavior of 8 common marmosets (4 males and 4 females consisting of relatives and non-relatives of various ages) housed in an 18-square-meter room filled with a variety of enrichment structures. 20 ring-shaped RFID antennas (FLEX-ID System by Phenovance LLC, Japan) were installed in key activity sites within the room (nest boxes, feeders, and water stations). The access information of each animal to these sites was recorded automatically, and the experimenter was able to extract and analyze the data remotely. Social behavior indices were explored based on continuous recording of basal activity patterns within the group. For cognitive assessment, an operant chamber with an RFID antenna (IntelliCage for rats by TSE systems, Germany) was placed inside the room. Tasks were designed remotely and executed automatedly, and each animal freely engaged in tasks. Results Based on access to various FLEX-ID antenna sites, we successfully tracked each animal's daily activity levels and movement patterns, and the progression of these events over several months. We derived the social relationship index from the encounter profiles at the nest boxes, which was in align with the caretakers' observations on how animals interacted with the others. For cognitive behavior analysis, we confirmed the adaptability of IntelliCage to test in group-housed common marmosets, with minor adjustments. After an automated pre-test training, individuals engaged in a fixed-ratio discrimination learning task, progressing at their own pace over several days where transitions to the next steps were determined by individual performance. We observed learning curves with notable individual variability in these animals. Discussion & Conclusion The automation of data acquisition is expected to enhance the reproducibility and accuracy of data interpretation. Here, we demonstrated that RFID technology-based behavioral monitoring enables a multi-index, continuous recording of group-housed animals, and can be flexibly adapted to an enriched housing setup which would be difficult with video-based analysis. This experimental approach is applicable for researching various psychiatric disease models and evaluating the drug effects more effectively.